Monitoring system and method of using same

This application is being filed as a PCT International Patent application on Apr. 2, 2020, in the name of Starkey Laboratories, Inc., a U.S. national corporation, applicant for the designation of all countries, and Justin R. Burwinkel, a U.S. Citizen, and Buye Xu, a Chinese Citizen, and Jason A. Galster, a U.S. Citizen, and Lauren Sculthorpe-Petley, a Canadian Citizen, and Peter J. Tetrick, a U.S. Citizen, and Sourav K. Bhunia, a U.S. Citizen, and Peggi S. Cross, a U.S. Citizen, and Martin F. McKinney, a U.S. Citizen, inventor(s) for the designation of all countries, and claims priority to U.S. Provisional Patent Application No. 62/828,766, filed Apr. 3, 2019, the contents of which are herein incorporated by reference in its entirety.

Activity monitoring and classification of patients and subjects can provide caregivers with real-time notice of medically-relevant events such as falls. For example, automated fall detection can detect whether a subject has fallen and send information regarding the fall to a caregiver or first responder. Maintaining postural control and preventing such falls are of importance for the elderly. Falls are the second leading cause of accidental or unintentional injury deaths worldwide and are especially prevalent in the elderly. Currently, individuals are often inadequately prepared to protect themselves from falls or other serious injuries as the onset of such events can come without perceptible warning. Further, maintaining postural equilibrium, i.e., prevention of a fall, involves stabilization of the body's center of mass during both self-initiated and externally triggered disturbances to postural stability during normal daily activities. Maintaining such equilibrium can be accomplished by limiting the motion of the center of mass within the base of support formed by and around the feet. Postural equilibrium is maintained through multisensory inputs. For example, loss of sensory input in the feet due to neuropathy can increase the risk of a fall, even though the necessary motor control for a corrective action of repositioning the feet can still be intact. Similarly, low vision or reduced ranges of hearing can prevent an individual from detecting hazards within their environment such that they can avoid them.

In general, the present disclosure provides various embodiments of a monitoring system. The monitoring system can include two or more sensors that are adapted to detect a characteristic of a subject of the system and generate data representative of the characteristic. The system can further include a controller operatively connected to the first and second sensors and adapted to determine statistics of an overall condition state of the subject over a monitoring time period based upon confirmation of one or more condition substates.

In one aspect, the present disclosure provides a monitoring system that includes first and second sensors each adapted to detect a characteristic of a subject of the system and generate data representative of the characteristic of the subject, where the characteristic includes at least one of a physiological characteristic or contextual information of the subject; and a controller operatively connected to the first and second sensors. The controller is adapted to receive data representative of a first characteristic of the subject from the first sensor, determine statistics for a first condition substate of the subject over a monitoring time period based upon the data received from the first sensor, receive data representative of a second characteristic of the subject from the second sensor, and determine statistics for a second condition substate of the subject over the monitoring time period based upon the data received from the second sensor. The controller is further adapted to compare the statistics of the first condition substate to the statistics of the second condition substate, confirm the first condition substate of the subject if the first condition sub state is substantially similar to the second condition substate, and determine the statistics of an overall condition state of the subject over the monitoring time period based upon the confirmed first condition substate.

In another aspect, the present disclosure provides a method that includes receiving data representative of a first characteristic of a subject from a first sensor, determining statistics for a first condition substate of the subject over a monitoring time period based upon the data received from the first sensor, receiving data representative of a second characteristic of the subject from a second sensor, determining statistics for a second condition substate of the subject over the monitoring time period based upon the data received from the second sensor. The method further includes comparing the statistics of the first condition substate to the statistics of the second condition substate, confirming the first condition substate of the subject if the first condition substate is substantially similar to the second condition substate, and determining the statistics of an overall condition state of the subject over the monitoring time period based upon the confirmed first condition substate.

These and other aspects of the present disclosure will be apparent from the detailed description below. In no event, however, should the above summaries be construed as limitations on the claimed subject matter, which subject matter is defined solely by the attached claims, as may be amended during prosecution.

In general, the present disclosure provides various embodiments of a monitoring system. The monitoring system can include two or more sensors that are adapted to detect a characteristic of a subject of the system and generate data representative of the characteristic. The system can further include a controller operatively connected to the first and second sensors and adapted to determine statistics of an overall condition state of the subject over a monitoring time period based upon confirmation of one or more condition substates.

Activity monitoring systems can be inaccurate due to individual variability among subjects and poor fitting devices. Further, monitoring systems that detect falls can be prone to falsely identifying such falls because of non-fall events such as sitting down. Such false alarms can unnecessarily escalate alarm and emergency responses from caregivers and first responders. Further, a subject of an automated fall detection system may become more fearful of falling, less forthcoming about their falls risk, or avoid using the fall detection system altogether. Fall detection accuracy is also important for a falls risk estimation model that can account for previous fall events detected by the system.

One or more embodiments of monitoring systems described herein can utilize two or more operatively connected sensors and devices to provide redundancy that can improve the accuracy of automated activity decisions. This redundancy results from either (a) use of different types of sensors, leading to similar conclusions that can be drawn (e.g., an accelerometer and a barometer can each indicate a sudden downward change in elevation) or (b) from a duplicity of similar sensors (e.g., an accelerometer disposed on a right ear of a subject corroborates a conclusion based upon measurements from an accelerometer disposed on a left ear of the subject).

electroencephalography (EEG) that can detect the onset of syncope (i.e., loss of consciousness), onset of a stroke or aneurysm, onset of a seizure, loss of alertness, onset of pain, etc.; EEG that can detect body movements that are not correlated with the subject's cognitive intent or that require greater degrees of cognitive load to perform (e.g., navigating obstacles, Parkinson's Disease, etc.); electrooculography (EOG) and videooculography (VOG) that can detect onset of nystagmus (i.e., indication of dizziness or intoxication); electrocardiography (EKG or ECG) that can detect onset of cardiac infarcts (i.e., heart attacks), palpitations, pericarditis, angina, etc.; blood pressure sensors that can detect the onset of orthostatic hypotension and orthostatic syncope; blood sugar level sensors that can detect the onset of hypoglycemia; blood oxygen level sensors that can detect the onset of hypoxia; carbon monoxide sensors that can corroborate symptoms of dizziness or syncope due to environmental factors; thermometers that can monitor changes in core body temperature or head temperature, or can detect an increased likelihood that a fall has occurred due to environmental factors; microphones that can detect an impact caused by a fall, vocal utterances related to distress, or corroborate presentation of the Tullio phenomenon (i.e., dizziness caused by loud sounds); ambient light sensors that can detect an increased likelihood that a fall has occurred due to various environmental factors; proximity sensors that can detect the presence of tripping hazards; and radio signal sensors that can communicate with nearby hazard beacons (either virtual or physical) or act as proximity sensors within a mesh network of other devices. One or more embodiments of monitoring systems described herein can utilize unique combinations of sensors for activity monitoring and fall detection that can enhance the classification of measurable events. Various inputs related to activity monitoring and fall detection may be detected either (a) leading up to a fall, (b) during a fall event, or (c) following a fall event. For example, one or more embodiments of monitoring systems described herein can utilize one or more of the following techniques and devices for at least one of activity monitoring or fall detection:

Output from various sensors can be used as raw data or classifications can be made based upon the sensor data. Further, one or more state models can utilize various statistical methods to determine which of the available inputs typically occur together given particular outcomes. These techniques can be applied, e.g., to activity classification and analysis and fall detection. Further, data from two or more sensors can assist medical professionals with classifying a fall event to determine if a subject has lost consciousness and the likelihood that a traumatic brain injury has occurred.

In addition, crowd-sourced data can be stored and accessed from the cloud. This information can assist in identifying specific locations or types of locations, events, groups of people, etc., where falls or other physical activities are likely to occur. For example, outdoor areas may include uneven ground that may cause falls or near-falls to occur. Further, falls may also commonly occur within indoor areas as well. Data regarding these outdoor and indoor areas can be automatically populated or manually populated by subjects to provide a warning to others. This information in turn can be utilized to form virtual hazard beacons as is described, e.g., in U.S. Patent Publication No. 2018/0233018A1, entitled FALL PREDICTION SYSTEM INCLUDING A BEACON AND METHOD OF USING SAME, the contents of which are herein incorporated by reference in its entirety.

1 FIG. 10 10 12 14 16 12 14 16 Any suitable system can be utilized for at least one of activity classification and fall detection. For example,is a schematic diagram of one embodiment of a monitoring system. The systemincludes a first sensor, a second sensor, and a controlleroperatively connected to the first and second sensors. As used herein, the term “operatively connected” means that an element or component can be connected to another element or component using any suitable technique or techniques such that information can be shared between such components. In one or more embodiments, one or both of the first sensorand second sensorcan be operatively connected to the controllerby a wire or cable, wirelessly using any suitable wireless protocol, optically, over the internet, a mesh network, etc.

10 10 100 10 10 16 2 FIG. The monitoring systemcan be utilized with any suitable device or application. For example, in one or more embodiments, the systemcan be associated with any suitable body-worn device, e.g., body-worn deviceof. In one or more embodiments, the systemcan be disposed in a room or building and be adapted to monitor one or more individuals proximate to the system using any suitable technique or techniques. In one or more embodiments, the systemcan utilize at least one of lidar, facial recognition software, camera systems, hazard beacons, Bluetooth-based directional arrival software, geotagging, radio frequency identification (RFID), light sensors, and the like to identify and track an individual located in the room or building and provide at least one of physiological data or contextual information regarding the individual to the controllerfor further processing as is described herein.

12 14 12 14 120 100 134 136 12 14 16 12 14 12 14 12 14 2 FIG. The sensors,can be disposed in any suitable location. In one or more embodiments, the sensors,can be a component of electronic componentsof the body-worn deviceof, e.g., sensors,. In one or more embodiments, the sensors,can be disposed outside of the housing of a hearing device and operatively coupled to the device and the controllerusing any suitable technique or techniques. In one or more embodiments, sensors,can be disposed within one or both ears and outside the ear of the subject. In one or more embodiments, the sensors,can be disposed within one or more devices worn by one or more individuals other than the subject of interest. In one or more embodiments, the sensors,can be disposed within one or more physical hazard beacons, e.g., as described in U.S. Patent Publication No. 2018/0233018 A1.

12 14 12 14 Each of the first and second sensors,, is adapted to detect a characteristic of a subject of the system and generate data representative of the characteristic of the subject. Any suitable characteristic or characteristics of the subject can be detected by the sensors,. In one or more embodiments, the characteristic includes at least one of a physiological characteristic or contextual information of the subject. As used herein, the term “physiological characteristic” refers to information regarding the subject's physiological state, e.g., at least one of a determined fall risk, inertial sensor data, heart rate information, blood pressure information, drug concentration information, blood sugar level, body hydration information, neuropathy information, blood oximetry information, hematocrit information, cortisol level, body temperature, age, sex, gait or postural stability attribute, vision, hearing, eye movement, neurological activity, head movement, or the like. In one or more embodiments, a physiological characteristic can include psychological data representative of a psychological state such as a fear of falling. A psychological state can, in one or more embodiments, be detected from data representative of the physiological characteristic (i.e., physiological data) or combinations of physiological characteristics such as heart rate, core body temperature, or cortisol levels. Further, in one or more embodiments, the physiological data can include, in part or in combination, one or more inputs provided by the subject in response to one or more queries.

10 Further, as used herein, the term “contextual information” refers to data representative of a context within which the subject is disposed or will be disposed at a future time. In one or more embodiments, data representative of the contextual information (i.e., contextual data) can include at least one of weather condition, environmental condition, sensed condition, location, velocity, acceleration, direction, hazard beacon, type of establishment occupied by the subject, camera information, or presence of stairs, etc. One or more hazard beacons can provide contextual data to the system. Such hazard beacons can include physical or virtual beacons as described, e.g., in U.S. Patent Publication No. 2018/0233018 A1.

12 14 10 12 14 12 14 12 14 Although depicted as including two sensors,, the systemcan include any suitable number of sensors, e.g., 1, 2, 3, 4, 5, or more sensors. Each of the first and second sensors,can include any suitable sensor or sensors. For example, each of the sensors,can include at least one of an accelerometer, barometer, gyroscope, heart rate sensor, blood pressure sensor, magnetometer, EOG sensor, EEG sensor, VOG sensor, EKG or ECG sensor, blood sugar sensor, blood oxygen sensor, light sensor, sweat sensor, pupillometry sensor, cerumen sensor, cortisol sensor, body temperature sensor, humidity sensor, air quality sensor, carbon monoxide sensor, microphone, proximity sensor, GPS, telecoil, magnetic sensor, or radio signal sensor, or combinations thereof. The sensors,can be adapted to detect any suitable characteristic of the subject, e.g., at least one of a physiological characteristic and contextual information of the subject. For example, the physiological characteristic can include at least one of body position, the position of specific body parts, eye movement, body temperature, heart rate, EEG, or skin impedance, or combinations thereof. Further, for example, the contextual information can include at least one of ambient temperature, moisture/humidity, sound, light intensity, terrain, elevation, ambient oxygen levels, pollutants, toxins, or carbon monoxide levels, or combinations thereof.

12 14 12 14 12 14 The first sensoris adapted to detect a first characteristic of the subject and generate data (e.g., physiological data or contextual data) representative of the first characteristic. Further, the second sensoris adapted to detect a second characteristic of the subject and generate data (e.g., physiological data or contextual data) representative of the second characteristic. The first and second characteristics of the subject can be any suitable characteristic, e.g., at least one of a physiological characteristic or contextual information of the subject. The first and second characteristics can include any suitable characteristic. The first characteristic detected by the first sensorcan be the same as or different from the second characteristic detected by the second sensor. For example, in one or more embodiments, the first characteristic detected by the first sensorcan be eye movement of the subject and the second characteristic detected by the second sensorcan be head movement of the subject.

12 14 16 16 16 16 12 14 Operatively connected to the first and second sensors,is the controller. The controllercan include any suitable controller. Further, the controllercan include any suitable electronic components or elements. The controlleris adapted to receive data representative of a first characteristic of the subject from the first sensor; determine statistics for a first condition substate of the subject over a monitoring time period based upon the data received from the first sensor; receive data representative of a second characteristic of the subject from the second sensor; determine statistics for a second condition substate of the subject over the monitoring time period based upon the data received from the second sensor; compare the statistics of the first condition substate to the statistics of the second condition substate; confirm the first condition substate of the subject if the first condition sub state is substantially similar to the second condition substate; and determine the statistics of an overall condition state of the subject over the monitoring time period based upon the confirmed first condition substate.

As used herein, the term “data” can include a single datum or a plurality of data values or statistics. The term “statistics” can include any appropriate mathematical calculation or metric relative to data interpretation, e.g., probability, confidence interval, distribution, range, or the like. Further, as used herein, the term “monitoring time period” means a period of time over which the first and second characteristics of the subject are measured and statistics for the first and second condition substates are determined. The monitoring time period can be any suitable length of time, e.g., 1 millisecond, 1 second, 10 seconds, 30 seconds, 1 minute, 10 minutes, 30 minutes, 1 hour, etc.

The first and second substates can each include any suitable substates that can be utilized to determine the overall condition state. In one or more embodiments, each of the first and second substates can include at least one of body posture, body motion, location, direction, subject's activity, or mental substate. The first condition substate can be the same as or different from the second condition substate.

16 16 Although described as including first and second substates, the controlleris operable to determine statistics for any suitable number of substates. For example, in one or more embodiments, the controlleris further adapted to determine statistics for a third condition substate of the subject over a second monitoring time period based upon the data received from the first sensor, and determine statistics for a fourth condition substate of the subject over the second monitoring time period based upon the data received from the second sensor. The third and fourth condition substates can be the same as or different from one or more of the first and second condition substates.

Any suitable technique or techniques can be utilized to determine statistics for the various substates, e.g., direct statistical analyses of time series data from the sensors, differential statistics, comparisons to baseline or statistical models of similar data and substates, etc. Such techniques can be general or individual-specific and represent long-term or short-term behavior. These techniques could include standard pattern classification methods such as Gaussian mixture models, clustering as well as Bayesian approaches, neural network models and deep learning.

16 Further, the controlleris adapted to compare statistics of a condition substate (e.g., the first condition substate) and the statistics of another condition substate (e.g., the second condition substate) using any suitable technique or techniques. For example, substate similarity and dissimilarity can be measured directly via standard statistical metrics such normalized Z-score, or similar multidimensional distance measures (e.g. Mahalanobis or Bhattacharyya distance metrics), or through similarities of modeled data and machine learning. These techniques can include standard pattern classification methods such as Gaussian mixture models, clustering as well as Bayesian approaches, neural network models, and deep learning.

16 Any suitable technique or techniques can be utilized by controllerto confirm a condition substate (e.g., the first condition substate) of the subject if the first condition sub state is substantially similar to the second condition substate. As used herein the term “substantially similar” means that, upon comparison, the sensor data are congruent or have statistics fitting the same statistical model, each with an acceptable degree of confidence. The threshold for the acceptability of a confidence statistic may vary depending upon the subject, sensor, sensor arrangement, substate, combination of substates, context, condition, etc. In one or more embodiments, the first condition substate can be confirmed by comparing sensor data or applying the sensor data to a statistical model for one or more substates. In one or more embodiments, the second condition substate can be confirmed if the second condition substate is substantially similar to the first condition substate. Any suitable technique or techniques can be utilized to confirm the second condition substate, e.g., the same techniques described herein regarding confirmation of the first condition substate.

The statistics of an overall condition state of the subject, over the monitoring time period, can be based upon the confirmed condition substate (e.g., first condition substate) by utilizing any suitable technique or techniques, e.g., standard pattern classification methods such as Gaussian mixture models, clustering, hidden Markov models, as well as Bayesian approaches, neural network models, and deep learning. As used herein, the term “overall condition state” refers to the predicted or realized status of the subject. Examples of an overall condition state may include an activity classification such as sitting, walking, running, jumping, falling intentionally, falling unintentionally, and the like.

16 16 The controllercan be adapted to determine the statistics of any suitable number of overall condition states of the subject over the monitoring time period. For example, in one or more embodiments, the controllercan further be adapted to compare the statistics of the third condition substate to the statistics of the fourth condition substate; confirm the third condition substate of the subject if the third condition substate is substantially similar to the fourth condition substate; and determine the statistics of a second overall condition state of the subject over the second monitoring time period based upon the confirmed third condition substate. Any suitable technique or techniques can be utilized to determine the statistics of the second overall condition state, e.g., the same techniques utilized to determine the statistics of the overall condition state.

16 The controllercan also be adapted to compare the statistics of the first overall condition state to the statistics of the second overall condition state. Any suitable technique or techniques can be utilized to compare the statistics of two or more overall condition states. e.g., Gaussian mixture models or Hidden Markov Models to learn and classify multiple sets of activities and/or stages of walking stability or falling. In one or more embodiments, ongoing activity can be continuously classified or monitored, and a fall prevention output can be generated based upon this classified activity.

16 In one or embodiments, the controlleris further adapted to determine one or more of a fall risk value and fall risk statistic based upon the determined overall condition state of the subject. Any suitable technique or techniques can be utilized to determine one or more of a fall risk value and fall risk statistic, e.g., the techniques described in U.S. Patent Publication No. 2018/0228404 A1, entitled FALL PREDICTION SYSTEM AND METHOD OF USING SAME; and U.S. Patent Application Ser. No. 62/785,295, entitled PREDICTIVE FALL EVENT MANAGEMENT SYSTEM AND METHOD OF USING SAME, the contents of which are herein incorporated by reference in their entirety.

16 In one or more embodiments, the controllercan be adapted generate a fall prevention output responsive to satisfaction of the fall condition. Any suitable fall prevention output can be provided that can assist in mitigating the risk of the subject falling, e.g., one or more of an audible alert, visual alert, or tactile alert provided to the subject. For example, the subject can be warned of the subject acting on a predicted behavior when the predicted risk associated with the predicted behavior crosses a threshold for alerting the subject. In one or more embodiments, the subject can be advised of suitable options or alternatives to acting on a particular predicted behavior to protect the subject from engaging in unnecessary risky behaviors. In one or more embodiments, the subject can be given directions or advice to mitigate their risks. In one or more embodiments, the use of virtual reality or augmented reality such as those described, e.g., in U.S. Patent Publication No. 2018/0233018A1 can be used to assist the subject in avoiding potential risks.

In one or more embodiments, the fall prevention output can include initiating a mitigation therapy or corrective measures e.g., a balance training regimen. The fall prevention output can further include controlling or modifying one or more features and actions of a mobility device, an assistive device, or an exoskeleton.

In one or more embodiments, the fall prevention output can include one or more modifications of an environmental context of the subject. Any suitable environmental contexts can be modified to satisfaction of a predicted fall condition. For example, the fall prevention output can include modifying, via mesh networks or the internet of things (IoT), e.g., the lighting within an area proximate to the subject or within an area where the subject is predicted to be. Similarly, the fall prevention output can include modifying the temperature, oxygen mix, humidity, or air quality of an area proximate to the subject or within an area where the subject is predicted to be. Further, in one or more embodiments, the fall prevention output can include transmission of one or more of the physiological data and the contextual data to one or more of a caregiver, a medical professional, a database, or the subject.

In one or more embodiments, the fall prevention output can also include adapting one or more of a monitoring system setting such that the system is more likely to indicate that a fall has occurred if the predicted fall risk statistics crossed a certain threshold leading up to the occurrence in question as described, e.g., as described in U.S. Patent Application Ser. No. 62/780,223, entitled HEARING ASSISTANCE SYSTEM WITH ENHANCED FALL DETECTION, the contents of which are herein incorporated by reference in its entirety. In one or more embodiments, the adaptation of system settings can be applied transitorily such that the system is only more likely to indicate that a fall has occurred for a period of seconds or minutes. In one or more embodiments, the adaptation of system settings can be applied over a longer period of time.

16 The controllercan further be adapted to determine a balance event of the subject if the statistics of an overall condition state indicates one or more of an unintentional loss of balance or postural stability and vestibular disturbance. Any suitable technique or techniques can be utilized to determine a balance event, e.g., the techniques described in U.S. Patent Application Ser. No. 62/785,295.

10 The systemcan be utilized with any suitable body-worn device for the subject that can be worn in any suitable location on the body of the subject, including but not limited to, a wearable hearing device such as headphones, a wrist-worn device such as a smartwatch, a patch disposed on any portion of a body of the subject, glasses, etc. The body-worn device can be implanted. Further, the body-worn device can, in one or more embodiments, be integrated into or otherwise disposed in a body part prothesis, mobility assistance device, or exoskeleton.

10 10 In one or more embodiments, the body-worn device can include a hearing assistance device such as behind-the-ear (BTE), in-the-ear (ITE), in-the-canal (ITC), or completely-in-the-canal (CIC) type hearing instrument. It is understood that behind-the-ear type hearing instruments can include devices that reside substantially behind the ear or over the ear. Such devices can include hearing instruments with receivers associated with the electronics portion of the behind-the-ear device or hearing instruments of the type having receivers in the ear canal of the subject. Such devices are also known as receiver-in-the-canal (RIC) or receiver-in-the-ear (RITE) hearing devices. In one or more embodiments, the body-worn device can include a cochlear implant (including its processor) or a bone-conduction or otherwise osseointegrated hearing device. It is understood that other body-worn devices not expressly stated herein can fall within the scope of the present subject matter. the systemcan be utilized with any suitable number of body-worn devices, e.g., two or more body-worn devices. For example, in one or more embodiments, the systemcan be utilized with a left hearing device that is adapted to be acoustically connected to the subject's left ear and a right hearing device that is adapted to be acoustically connected to the subject's right ear. In one or more embodiments, the left hearing device can electrically communicate with the right hearing device using any suitable technique or techniques.

2 FIG. 1 FIG. 1 FIG. 100 100 110 120 120 120 122 16 124 126 128 130 134 136 12 14 124 126 128 130 134 136 122 100 The body-worn device can include any suitable electronic components or circuitry. In one or more embodiments, the body-worn device can include hearing assistance components. For example,is a schematic diagram of one embodiment of a body-worn device. The deviceincludes a housingand hearing assistance componentsdisposed within the housing. Hearing assistance componentscan include any suitable device or devices, e.g., integrated circuits, power sources, microphones, receivers, etc. For example, in one or more embodiments, the componentscan include a controller(e.g., controllerof), a microphone, a receiver(e.g., speaker), a power source, an antenna, and one or more sensors,(e.g., first and second sensor,of). The microphone, receiver, power source, antenna, and sensors,can be electrically connected to the controllerusing any suitable technique or techniques. In one or more embodiments, the body-worn device can be adapted to provide the subject with auditory information, feedback, or guidance. In one or more embodiments, the devicecan be adapted to accept verbal or gestural input from the subject. In one or more embodiments, encephalography (EEG) sensor data can be utilized to infer e.g., the subject's intentions.

122 100 16 10 122 100 122 122 1 FIG. Any suitable controllercan be utilized with the body-worn device, e.g., the same controller or controllers described herein regarding controllerof systemof. For example, the controllercan be adapted to employ programmable gains to adjust the output of the deviceto a patient's particular hearing impairment. The controllercan be a digital signal processor (DSP), microprocessor, microcontroller, other digital logic, or combinations thereof. The processing can be done by a single processor or can be distributed over different devices. The processing of signals described herein can be performed using the controlleror over different devices.

122 132 122 122 124 126 In one or more embodiments, the controlleris adapted to perform instructions stored in one or more memories. Various types of memory can be used, including volatile and nonvolatile forms of memory. In one or more embodiments, the controlleror other processing devices execute instructions to perform a number of signal processing tasks. Such embodiments can include analog components in communication with the controllerto perform signal processing tasks, such as sound reception by the microphone, or playing of sound using the receiver.

122 16 1 FIG. In general, digital hearing devices include a controller or processor. In such devices, programmable gains our audiometric filters can be employed to adjust the hearing device output to a subject's particular hearing impairment or preferences. The controller(and controllerof) can be a digital signal processor (DSP), microprocessor, microcontroller, other digital logic, or combinations thereof. The processing can be performed by a single processor or can be distributed over different devices. The processing of signals referenced in this application can be performed using the processor or other different devices. Processing can be done in the digital domain, the analog domain, or combinations thereof. Processing can be done using subband processing techniques. Processing can be done using frequency domain or time domain approaches. Some processing can involve both frequency and time domain aspects. For brevity, in some examples drawings can omit certain blocks that perform frequency synthesis, frequency analysis, analog-to-digital conversion, digital-to-analog conversion, amplification, buffering, and certain types of filtering and processing. In various embodiments, the processor is adapted to perform instructions stored in one or more memories, which can or cannot be explicitly shown. Various types of memory can be used, including volatile and nonvolatile forms of memory. In various embodiments, the processor or other processing devices execute instructions to perform a number of signal processing tasks. Such embodiments can include analog components in communication with the processor to perform signal processing tasks, such as sound reception by a microphone, or playing of sound using a receiver (i.e., in applications where such transducers are used). In various embodiments, different realizations of the block diagrams, circuits, and processes set forth herein can be created by one of skill in the art without departing from the scope of the present subject matter.

120 124 122 124 120 124 110 110 124 The hearing assistance componentscan also include the microphonethat is electrically connected to the controller. Although one microphoneis depicted, the componentscan include any suitable number of microphones. Further, the microphonecan be disposed in any suitable location within the housing. For example, in one or more embodiments, a port or opening can be formed in the housing, and the microphonecan be disposed adjacent the port to receive audio information from the subject's environment.

124 124 122 122 124 Any suitable microphonecan be utilized. In one or more embodiments, the microphonecan be selected to detect one or more audio signals and convert such signals to an electrical signal that is provided to the controller. Although not shown, the controllercan include an analog-to-digital convertor that converts the electrical signal from the microphoneto a digital signal.

122 126 126 122 110 126 112 114 110 126 112 114 216 110 Electrically connected to the controlleris the receiver. Any suitable receiver can be utilized. In one or more embodiments, the receivercan be adapted to convert an electrical signal from the controllerto an acoustic output or sound that can be transmitted from the housingto the subject. In one or more embodiments, the receivercan be disposed adjacent an openingdisposed in a first endof the housing. As used herein, the term “adjacent the opening” means that the receiveris disposed closer to the openingin the first endthan to a second endof the housing.

128 122 120 128 128 120 128 The power sourceis electrically connected to the controllerand is adapted to provide electrical energy to the controller and one or more of the other hearing assistance components. The power sourcecan include any suitable power source or power sources, e.g., a battery. In one or more embodiments, the power sourcecan include a rechargeable battery. In one or more embodiments, the componentscan include two or more power sources.

120 130 130 110 130 120 The componentscan also include the optional antenna. Any suitable antenna or combination of antennas can be utilized. In one or more embodiments, the antennacan include one or more antennas having any suitable configuration. For example, antenna configurations can vary and can be included within the housingor be external to the housing. Further, the antennacan be compatible with any suitable protocol or combination of protocols. In one or more embodiments, the componentscan also include a transmitter that transmits electromagnetic signals and a radio-frequency receiver that receives electromagnetic signals using any suitable protocol or combination of protocols.

100 100 100 100 100 For example, in one or more embodiments, the body-worn device(or any other body-worn device described herein) can be connected to one or more external devices using, e.g., Bluetooth, Wi-Fi, magnetic induction, mesh networks, etc. For example, in one or more embodiments, the body-worn devicecan be wirelessly connected to the Internet using any suitable technique or techniques. Such connection can enable the body-worn deviceto access any suitable databases, including medical records databases, cloud computing databases, location services, etc. In one or more embodiments, the body-worn devicecan be wirelessly connected utilizing the Internet of Things (IoT) such that the hearing device can communicate and share data with, e.g., one or more hazard beacons, one or more cameras disposed in proximity to the subject, motion sensors, room lights, air conditioning and heating controllers, etc. Further, in one or more embodiments, the body-worn devicecan access weather information via the Internet or a mesh network using any suitable technique or techniques such that the subject can be informed of potentially hazardous weather conditions.

100 134 136 134 136 100 134 136 12 14 10 134 136 134 136 134 136 122 1 FIG. In one or more embodiments, the body-worn devicecan include the first sensorand the second sensor. Although depicted as including two sensors,, the body-worn devicecan include any suitable number of sensors, e.g., 1, 2, 3, 4, 5, or more sensors. The sensors,can include any suitable sensor or sensors, e.g., the same sensors described herein regarding sensors,of systemof. The first sensorcan include the same sensor as the second sensor. In one or more embodiments, the first sensorincludes a sensor that is different from that of the second sensor. The sensors,can be operatively connected to the controllerusing any suitable technique or techniques.

134 100 136 100 12 14 10 134 136 134 136 1 FIG. In one or more embodiments, the first sensoris operatively connected to the body-worn deviceand adapted to detect a first characteristic of the subject and generate data (e.g., physiological data or contextual information) representative of the first characteristic. In one or more embodiments, the second sensoris operatively connected to the body-worn deviceand adapted to detect a second characteristic of the subject and generate data (e.g., physiological data or contextual information) representative of the second characteristic. The first and second characteristics can include any suitable characteristic, e.g., the same characteristic or characteristics described herein regarding sensors,of systemof. The first characteristic detected by the first sensorcan be the same as or different from the second characteristic detected by the second sensor. For example, in one or more embodiments, the first characteristic detected by the first sensorcan be eye movement of the subject and the second characteristic detected by the second sensorcan be head movement of the subject.

1 FIG. 10 10 16 12 14 10 10 Returning to, the monitoring systemcan be utilized to receive input information and determine the statistics of one or more overall condition states of the subject or the monitoring time period. In one or more embodiments, the systemcan be utilized to receive input information from any suitable source to determine the statistics of the overall condition state. The input information can be provided using any suitable sensor, device, or database. For example, the input information can be provided to the controllerby the sensors,, the system, manually by one or more of the subject, a caregiver, and a medical professional, or obtained from other systems via wired or wireless connections to system.

10 10 10 10 Further, the monitoring systemcan provide any suitable outputs that can be based on the overall condition state or states. Any suitable output or outputs can be provided by the system, e.g., notifications, reports, IoT triggers (e.g., activating room lighting), treatments to the subject of the system, etc. In one or more embodiments, the systemcan be utilized to detect head impact, check with the subject for consciousness, and inform one or more of the subject, caregiver, and medical professional of the detection of a head impact and level of consciousness of the subject.

10 10 12 14 As mentioned herein, the systemcan utilize any suitable sensor configuration. For example, the systemcan be utilized to detect and confirm Tullio's Phenomenon in a subject. Tullio's Phenomenon is characterized by loud external sounds or the subject's own voice causing dizziness that is evidenced with nystagmus eye movement. To detect and confirm this Phenomenon, the sensors,of the system can include one or more of a microphone, eye movement sensor (e.g., electronystagmography (ENG), or videonystagmography (VNG) using a camera in smartglasses or on smartphone, etc.), own-voice sensor (e.g., accelerometer in the ear, microphone array, etc.), and an IMU sensor adapted to detect postural adjustments in response to the sound detected at the microphone.

10 12 14 Further, in one or more embodiments, the systemcan be utilized to detect and confirm a loss of balance of the subject. The sensors,can include one or more of an IMU sensor adapted to detect rapid stepping, sway, short-term change in walking stride/pattern, etc. (preceding a fall or near fall event), an EEG sensor (EEG can suggest that the brain is attempting to make the reflexive movements (e.g., rapid stepping) needed to recover from lost balance (preceding a fall or near fall event)), an eye movement sensor adapted to detect sway or abnormal stepping when there is an indication of nystagmus/loss of balance (preceding a fall or near fall event), and a microphone adapted to detect (acoustically) the presence of rapid stepping and when there is an indication of nystagmus/loss of balance (preceding a fall or near fall event).

10 Further, differentiation between reflexive and voluntary control of posture and movement can be detected and confirmed by the systemutilizing an IMU sensor combined with an EEG sensor. In such an embodiment, EEG data from the EEG sensor can suggest that the brain is attempting to make certain muscle movement. The system can determine if such EEG data corresponds with subject activity that's being measured by the IMU sensor. A mismatch between EEG data and IMU data can suggest a neurological problem.

10 The systemcan also be utilized to detect and confirm a hypotensive event. For example, heart rate sensors can be used to detect heart rate variability to predict an orthostatic hypotensive event. Such hypotensive event can be confirmed by data from a skin temperature sensor to detect vasoconstrictor failure due to inadequate release of norepinephrine in neurogenic orthostatic hypotension.

10 200 200 10 10 200 16 10 200 16 200 16 200 3 FIG. 1 FIG. 1 FIG. 1 FIG. As described herein, the monitoring systemcan utilize any suitable technique or techniques to determine statistics of one or more overall condition states of a subject of the system. For example,is a flowchart of one embodiment of a methodfor determining statistics of an overall condition state of a subject. The methodcan be implemented using any suitable device or system, e.g., the monitoring systemof. Although described in reference to systemof, the methodcan be utilized with any suitable device or system. In one or more embodiments, controllerof systemofcan be adapted to perform method. Further, the controllercan be adapted to perform the methodthrough firmware, software, etc. In one or more embodiments, the controllercan be or include an application-specific integrated circuit (ASIC) that includes the method.

200 202 204 202 12 14 10 202 202 202 Methodincludes receiving inputsatfor signal/data preparation. Any suitable inputscan be received. In one or more embodiments, the inputs can be data representative of a first characteristic of a subject from the first sensor. Further, in one or more embodiments, data representative of a second characteristic of the subject can be received from the second sensoras one or more inputs. In one or more embodiments, one or both of the first characteristic data and second characteristic data can include physiological data representative of a physiological characteristic of the subject of the systemover a monitoring time period. Further, one or both of the first characteristic data and second characteristic data can include contextual information representative of contextual information of the subject over the monitoring time period. Medical/health reports regarding the subject can also be provided as inputs. Any suitable databases containing, e.g., calendar data, reminder data, and data relative to actions of an artificial intelligence can also be provided as inputs. Further, group data from the cloud can also be provided as inputs at.

204 202 204 Such data can be prepared atutilizing any suitable technique or techniques. In one or more embodiments, data from inputscan be, e.g., filtered, time stamped, assigned metadata memory locations, and combinations thereof, at.

206 12 208 202 At, statistics for a first condition substate of the subject can be determined over a monitoring time period based upon the data received from the first sensorand stored at. Any suitable technique or techniques can be utilized to determine the statistics for the first condition substate. For example, one or more pattern recognition machine learning techniques such as Hidden Markov Models (HMMs), Bayesian models, non-linear regressions, decision trees, support vector machines, neural networks, etc. can be applied to input datato predict future physiological and contextual states based upon, at least in part, one or more of present and historical data. During a learning stage, prepopulated group data or laboratory data can be utilized to form the basis of training one or more pre-trained models. As the system collects data about the subject over time, one or more of these models can be updated using any suitable training techniques.

206 214 210 216 210 For example, a particular model atcan be trained or continuously adapted atto learn the effects that a specific medication has on an individual, over time, based on data from a group of individuals. The model can be further updated, for a specific subject, as specific data relative to the severity and timing of any observable changes to the subject that is collected. In this example, an observable change can relate to one or more characteristics of the subject's gait and postural stability that can be observed or monitored using any suitable sensor arrangement. In one or more embodiments, the time at which the subject will ingest a particular medication can be predicted based upon any suitable contextual data e.g., historical, physiological, or historical contextual data, scheduled reminders, push notifications, location information, etc. Further, a particular model atcan be trained or continuously adapted atbased on the subject's contextual information such as historical environment and location data, over time, to recognize patterns such that a prediction of the subject's future location can be made at. For example, historical fall event and balance event data from one or more of the individual or a group of individuals, at a given location, can further inform the risk factor associated with the location the individual is predicted to be at in a future time.

210 14 212 12 14 Statistics for a second condition substate of the subject can be determined atover a monitoring time period based upon the data received from the second sensorand stored at. Any suitable technique or techniques can be utilized to determine the statistics for the second condition substate, e.g., the same techniques utilized to determine statistics for the first condition substate. Although not shown, additional substates can be determined based upon additional data from the first and second sensors,or one or more additional sensors.

206 214 202 The first condition substatecan be updated atbased upon one or more of the inputs. For example, changes in gait such as increased postural sway, irregular stride, lower foot clearance, and decreased walking speed can be detected, using any suitable means, and can indicate that a new lower limb limitation can increase the individual's risk for stumbling over uneven walking paths.

210 216 202 210 The second substate modelcan also be updated atbased upon one or more of the inputsand the contextual risk factor values or statistics. For example, changes in weather conditions or geographical locations of the subject can be factored into the contextual factor modelsusing any suitable technique or techniques.

218 218 At, statistics of the first condition substate can be compared to the statistics of the second condition substate using any suitable technique or techniques. In one or more embodiments, statistics of the third condition substate can be compare to the statistics of the fourth condition substate atusing any suitable technique or techniques.

218 218 Also at, the first condition substate of the subject can be confirmed using any suitable technique or techniques. For example, the first condition substate can be confirmed if the first condition substate is substantially similar to the second condition substate. Further the third condition substate of the subject can be confirmed atusing any suitable technique or techniques. For example, the third condition substate can be confirmed if the first condition substate is substantially similar to the forth condition substate.

218 218 218 The statistics of one or more overall condition states of the subject over the monitoring time period based upon the confirmed one or more condition substates can be determined at. For example, the statistics of an overall condition state of the subject over the monitoring time period based upon the confirmed first condition substate can be determined at. Any suitable technique or techniques can be utilized to determine the statistics of the overall condition state. Further, the statistics of the second overall condition state of the subject over the second monitoring time period based upon the confirmed third condition substate can also be determined atusing any suitable technique or techniques.

218 200 12 14 In one or more embodiments, one or more of a fall risk value and fall risk statistic based upon the determined overall condition state of the subject can be determined atusing any suitable technique or techniques. For example, in one or more embodiments, a risk or probability value of a fall can be determined that can be based on a predetermined formula or formulas that can be derived from experimental data. The formula can also be entirely learned or modified through various machine learning approaches. For example, when a fall event is detected, the methodcan send postural data collected before the event by one or more sensors,, e.g., to a cloud server. In one or more embodiments, data from the subject and other subjects can be used to train a regression model or deep neural network to estimate the risk of a fall for an individual subject. One or more of the sub-models can be generated from analytics or machine learning of larger group data using any suitable technique or techniques, e.g., regression, steady-state, Bayesian, classification trees, Volterra, support vector machine, Gaussian mixture, neural network techniques, and combinations thereof.

10 12 14 The sub-models can provide an estimate or probability of the a given substate being an accurate reflection of the user of the subject and learn the subject's norms regarding motion patterns and health/physiological information. Inputs for generating the sub-models can either be obtained based on clinical evaluations and medical history or be learned by the monitoring systemfrom one or more inputs provided by various types of sensors, e.g., sensor,, and responses to queries.

For example, motion patterns of the subject and changes to such patterns can be estimated and monitored based on the outputs from one or more of an inertial measurement unit (IMU) sensor, GPS sensor barometer, magnetometer, EEG sensor, camera, etc. The motion of the subject can include sway amplitude and speed while walking, speed and trajectory when sitting down or standing up, speed and radius when turning, stride length, symmetry and variance, frequency of walks, length or distance of walks, reaction speed, etc.

200 In one or more embodiments, one or more of periodic assessments of functional balance ability, muscle strength, perceived balance efficacy, fear of falling, or functional reaction speed of the subject can be performed by the methodand queried to the subject or a third party. In one or more embodiments, one or more of results of periodic functional balance ability, muscle strength, perceived balance efficacy, fear of falling, or functional reaction speed of the subject can be entered into the manually by either the subject or a third party. As an illustrative example, the results of e.g., the Timed Up and Go (TUG) test can be calculated either by the instrumented body-worn device or entered by the individual or a third-party observer.

218 12 14 In one or more embodiments, physiological data that can be provided as inputs to the sub-models atinclude heart rate, blood pressure, blood sugar, blood oxygen, core body temperature, etc., and can be monitored utilizing any suitable sensor or sensors,. All such inputs and how they change over time can be monitored and used to estimate whether a fall condition is satisfied (i.e., how prone the subject is to a fall).

218 206 12 14 12 14 For example, one of the sub-models atcan evaluate postural stability (i.e., displacement of the head of the subject in three dimensions) of the subject to determine a fall risk value based on monitoring of the subject at. Any suitable sensor or sensors,can be utilized to determine postural stability, e.g., one or more of an accelerometer, gyroscope, microphone, barometer, optical sensor, and bioelectrical sensor. In one or more embodiments, the sensors,can include an accelerometer and a gyroscope as the primary sensors for postural balance and fall-risk monitoring and the other sensors can be secondary sensors. For example, a secondary sensor can include a microphone that can be used for detecting foot-falls or a fall event. Further, a barometer can be used to detect stair climbing. In addition, an optical sensor can be used for measuring heart rate and other biosignals. A bioelectric sensor can be used for monitoring electro-, cardio-, encephalo-, occulo-, and myo-graph signals from any location on the head and body of the subject.

In general, there can be multiple activities and postures during which one may fall down, most commonly walking and standing, transitions between postures such as movement between standing and sitting, etc. Further, there can be identifiable physiological events that precede the fall, such as postural hypotension.

12 14 One or more physiological sensors,can be employed to identify a “prodrome” of a postural instability. Some possible techniques of using this sensor information for this purpose can be used individually or in combination.

12 14 12 14 16 12 14 202 For example, in one or more embodiments, the sensors,can include one or more of an accelerometer and a gyroscope. Signals form the sensors,,can be used to compute and monitor a deviation from a stable position and a velocity with which such deviation takes place. In one or more embodiments, the controllercan utilize the signal inputs from the sensors,to generate a measure of postural stability. Such postural stability can be included in one or more physiological data inputs at. Postural stability can be recorded during normal daily activities, including standing, walking, and climbing stairs. Postural stability can also be recorded during structed activities and functional tests, e.g., during performance of a TUG test or the like. A threshold of normal stability can be established based on clinical postural stability testing or during a subject-initiated initialization involving one or more of these activities. Measurements in case of a recorded fall can be used to adjust the threshold, if appropriate.

Acceleration of the head of the subject while walking is complex, with the most prominent feature in the unprocessed accelerometer signal being that of the footfall. Adding to this complexity can be stabilization of the head by the neck. Footfall signals can be diminished by neck stabilization but still can be detectable. Vestibular-ocular reflexes can also be measured as the eye will attempt to stabilize the individual's visual field with each step. In one or more embodiments, head oscillation in three dimensions (antero-posterior (AP), lateral, and vertical) can be measured. Components of the displacement and the velocity in each dimension can be computed as measures of the postural stability. Although generally correlated and constrained by the body, the head can move relatively independently, which introduces artifacts. To mitigate these artifacts, in one or more embodiments, the velocity and displacement of the head oscillation are computed only when the pitch, yaw and/or roll motions of the head a slower than some predefined thresholds. Artifacts related to head movements can also be mitigated, by the controller, through the integration of sensor inputs of body-worn sensors placed on the chest, trunk, waist, etc. The values or statistics can depend upon the speed and type of body movement.

16 16 In one or more embodiments, the controllercan be adapted to determine a fall condition by measuring a maximum displacement between a longitudinal axis of the subject and a normal to the earth's surface as a function of time. Further, in one or more embodiments, the controllercan be adapted to determine the fall condition by measuring a maximum velocity of displacement between a longitudinal axis of the subject and a normal to the earth's surface.

Fall risk thresholds related to safe postural stability or limits of stability can be established by balance testing in a clinical setting or by subject-conducted self-directed tests. A fall risk signal or other fall risk output can be generated based on single or multiple threshold crossings.

Parameters of postural stability, i.e., balance metrics, and fall risk values or statistics can be of interest to one or more of the subject, caregivers such as the family members, and medical professionals. Balance metrics and fall risk values or statistics can be monitored e.g., daily or hourly and transmitted to various parties. The system can continuously monitor the subject, and once a fall risk threshold is exceeded, a fall risk output such as a discrete audio alert can be provided to the subject.

12 14 10 In laboratory conditions, head worn IMU sensors can be utilized to characterize small motions (e.g., sway) that can be important for balance evaluation. The orientation of the IMU sensors, however, is highly controlled and well calibrated in the laboratory. In practice, when subjects are wearing two hearing devices, proper alignment of the IMU sensors at each side of the head is desired. Any suitable technique or techniques can be utilized to align the sensors,in both left and right hearing devices of the system, e.g., the techniques described in U.S. patent application Ser. No. 15/331,130, filed Oct. 21, 2016, and entitled HEAD RELATED TRANSFER FUNCTION INDIVIDUALIZATION FOR HEARING DEVICE, the contents of which are herein incorporated by reference in its entirety. In one or more embodiments, a technique can be utilized to compensate for the orientation mismatch between two hearing devices so that the IMU sensors on both sides of the head can be collaboratively aligned with the head orientation and used to derive postural stability information.

12 14 In one or more embodiments, the fall risk value based upon postural stability can be determined by first detecting that the subject is walking. One or more artifacts from the sensors,caused by foot impact can be filtered out using any suitable technique or techniques. Postural stability can be determined using any suitable technique or techniques. Velocity components of such postural stability can also be determined using any suitable technique or techniques. In one or more embodiments, the fall risk value can be based upon walking speed, distance walked, frequency of walks, duration of walks, frequency of successful postural transitions, speed of postural transitions, or the like and other activity classifications, and combinations thereof.

12 14 A composite sensitivity parameter of the contribution of the sensors,(e.g., one or more accelerometers) to the overall fall risk value or statistic can be determined using any suitable technique or techniques. In one or more embodiments, the sensitivity of the fall risk value or statistics to an amplitude of the postural stability can be determined using, e.g., one or more of a subject input after a near-fall event, a balance study, and fall detection. The sensitivity of the fall risk value or statistics to the stability velocity at a pre-determined postural stability can be determined using, e.g., one or more subject inputs after a near-fall event, a balance test, fall detection or the like. Further, the sensitivity of the fall risk value to a statistically determined combination of the postural stability and the stability velocity can also be determined.

In one or more embodiments, postural stability, sway velocity and other posture, walking and fall-related information can be routinely transmitted to healthcare professionals. The subject's posture while standing and walking, actual fall events, and subject-indicated near-fall events can also be transmitted to healthcare professionals.

If the fall risk value or statistics crosses a fall risk threshold, then an alert can be sent to one or more of the subject, caregiver, and medical professional. Such alerts can include instructions for how to prevent a fall from occurring.

12 14 12 14 In one or more embodiments, sensors,having one or more accelerometers can be placed in both ears of the subject. Acceleration of the mid-point between the two ears, as opposed to that of one ear, can be calculated to determine postural stability. Further, false positives of fall detection can be reduced by ensuring both sensors,follow the same nominal motion pattern. In addition, head rotation around the vertical axis i.e., the yaw, can also be determined and utilized to calculate the fall risk value.

10 12 14 10 10 12 14 12 14 In one or more embodiments, a fall condition can be determined by measuring eye movement of the subject. For example, the monitoring systemcan detect eye movements and compare such eye movements to a baseline to determine whether a vestibular event is occurring that can increase the risk of fall. The sensors,of the monitoring systemcan include one or more eye movement sensors. In one or more embodiments, the systemcan also include one or more sensors,that can measure head movement of the subject. Data from such head movement sensors,can be utilized to correlate with eye movement sensor data to determine the risk of a fall. Any suitable fall prediction system or device can be utilized to measure eye movement of a subject, e.g., the devices described in U.S. Pat. No. 9,167,356, issued Oct. 20, 2015, and entitled ELECTROOCULOGRAM AS A CONTROL IN A HEARING ASSISTANCE DEVICE, the contents of which are herein incorporated by reference in its entirety.

10 For example, in one or more embodiments, one or more inputs can be provided by one or more of the subject, the caregiver, and the physician. For example, one or more inputs can be provided by the subject in response to one or more queries provided, e.g., by the system, the caregiver, or the physician.

218 220 222 224 Outputs from the state modelscan be stored at. Such outputs can include data that includes risk-level scaling and trends. Ata determination, for a future time, of whether an overall condition state satisfies one or more criteria can be made using any suitable technique or techniques. For example, each overall condition state can be assigned a confidence level using any suitable technique or techniques. The confidence level of each overall condition state can be compared to one or more thresholds to determine whether one or more outputs or actions are triggered at.

224 224 222 218 10 16 222 16 222 16 222 For example, reinforcement learning techniques can be applied atto select the best possible output behavior strategy in a given situation. Reinforcement learning models can consider one or more of a predicted physiological risk factor, predicted contextual risk factor, historical physiological risk factor, historical contextual factor, predicted fall event, historical fall event, predicted balance event, historical balance events, and the like. In one or more embodiments, the reinforcement learning model can further consider the statistics of an overall condition state associated with historical data to optimize one or more outputs at. In one or more embodiments, determining whether the overall condition state threshold is satisfied atincludes applying a future fall risk value to a fall risk model. Such fall risk model can be developed from outputs from the fall risk prediction state models. In one or more embodiments, the systemcan include several models: a physiological model, a contextual model, and a fall risk model. The fall risk model can use the output (e.g., state) of the physical model and contextual model as inputs to the fall risk model. The future fall risk value can be based upon the one or more future physiological states or contextual states. In one or more embodiments, the controllercan be adapted to determine at, for each of a plurality of future times, whether an anticipated fall condition is satisfied based upon one or more of the physiological data and contextual data. Further, in one or more embodiments, the controllercan be adapted to determine at, for a plurality of future times, whether an anticipated fall condition is satisfied based upon a combination of the physiological data and the contextual information. In one or more embodiments, the controllercan be adapted to determine at, for a plurality of future times, whether the fall condition is satisfied by determining a future fall risk value based upon the one or more future physiological states or contextual states.

226 222 222 224 Data related to thresholds can be stored atand utilized atto determine whether the overall condition state threshold has been satisfied atand also utilized to generate outputs at.

222 200 224 228 230 232 If the overall condition state threshold has been met at, then the methodproceeds to, where one or more outputs responsive to satisfaction of one or more overall condition state thresholds can be provided to the subject. The output can include one or more outputs described herein. For example, a fall prevention output can be provided to the subject atthat includes one or more alerts that are provided to one or more of the subject, caregiver, and medical professional for proper diagnosis and treatment. Further, for example, an output can include initiating mitigation therapy at. Such mitigation therapy can include initiating a balance training regimen or corrective measure as described, e.g., in U.S. Patent Publication No. 2018/0317837A1, entitled HEARING ASSISTANCE DEVICE INCORPORATING VIRTUAL AUDIO INTERFACE FOR THERAPY GUIDANCE, the contents of which are herein incorporated by reference in its entirety. In one or more embodiments, the fall prevention output can include delivery or modification of a therapy or initiation of an intervention, e.g., activation of an exoskeleton worn by the subject. Further, the fall prevention output can include generating an IoT action atsuch as sending one or more signals to one or more IoT devices proximate to the subject to help prevent the fall or otherwise protect the subject from injury, increasing an intensity of the ambient light for the subject's environment, etc. In one or more embodiments, an output can include a fall prevention output if the fall risk value or fall risk statistic crosses a fall risk threshold.

222 224 234 234 214 216 218 A particular instance of when the overall condition state threshold is satisfied atand which outputs are generated atcan be stored at. Such data that is stored atcan be utilized to updated one or both of the physiological factor models ator the contextual factor models at. Further, such data of a fall condition being met can be utilized to modify one or more of the fall risk prediction sub-models at.

222 200 236 If the overall condition state threshold has not been satisfied at, then the methodproceeds to a wait condition at.

All headings provided herein are for the convenience of the reader and should not be used to limit the meaning of any text that follows the heading, unless so specified.

The terms “comprises” and variations thereof do not have a limiting meaning where these terms appear in the description and claims. Such terms will be understood to imply the inclusion of a stated step or element or group of steps or elements but not the exclusion of any other step or element or group of steps or elements.

The words “preferred” and “preferably” refer to embodiments of the disclosure that may afford certain benefits, under certain circumstances; however, other embodiments may also be preferred, under the same or other circumstances. Furthermore, the recitation of one or more preferred embodiments does not imply that other embodiments are not useful, and is not intended to exclude other embodiments from the scope of the disclosure.

In this application, terms such as “a,” “an,” and “the” are not intended to refer to only a singular entity but include the general class of which a specific example may be used for illustration. The terms “a,” “an,” and “the” are used interchangeably with the term “at least one.” The phrases “at least one of” and “comprises at least one of” followed by a list refers to any one of the items in the list and any combination of two or more items in the list.

The phrases “at least one of” and “comprises at least one of” followed by a list refers to any one of the items in the list and any combination of two or more items in the list.

As used herein, the term “or” is generally employed in its usual sense including “and/or” unless the content clearly dictates otherwise.

The term “and/or” means one or all of the listed elements or a combination of any two or more of the listed elements.

As used herein in connection with a measured quantity, the term “about” refers to that variation in the measured quantity as would be expected by the skilled artisan making the measurement and exercising a level of care commensurate with the objective of the measurement and the precision of the measuring equipment used. Herein, “up to” a number (e.g., up to 50) includes the number (e.g., 50).

Also herein, the recitations of numerical ranges by endpoints include all numbers subsumed within that range as well as the endpoints (e.g., 1 to 5 includes 1, 1.5, 2, 2.75, 3, 3.80, 4, 5, etc.).

All references and publications cited herein are expressly incorporated herein by reference in their entirety into this disclosure, except to the extent they may directly contradict this disclosure. Illustrative embodiments of this disclosure are discussed and reference has been made to possible variations within the scope of this disclosure. These and other variations and modifications in the disclosure will be apparent to those skilled in the art without departing from the scope of the disclosure, and it should be understood that this disclosure is not limited to the illustrative embodiments set forth herein. Accordingly, the disclosure is to be limited only by the claims provided below.